Are any functionally mature cells of medullary phenotype located in the thymus cortex?

Experiments were undertaken to test if thymocytes of "mature" or "medullary" phenotype were restricted to the medullary area of the thymus. A calculation based on direct cell counts on serial sections indicated that 11.5% of adult male CBA thymic lymphoid cells were within the medullary zone. Since only 3-4% of thymocytes were cortisone resistant, the majority of thymocytes within the medulla were, like cortical thymocytes, cortisone sensitive. A series of cell surface antigenic markers, used alone or in pairs, suggested that 13-15% of thymocytes were of medullary phenotype, somewhat more than the number of thymocytes actually present in the medulla. However, much of this discrepancy could be explained by differential death of cortical cells during isolation and staining, and by the existence in the cortex of a subpopulation of early blast cells which shared some, but not all markers with medullary thymocytes. A direct test for mature or medullary phenotype cells in the cortex involved selective transcapsular labeling of outer-cortical cells with fluorescent dyes, followed by multiparameter immunofluorescent analysis of the 10% labeled population. Outer-cortical thymocytes included some cells (mainly early blasts) sharing some markers with medullary thymocytes, but very few (less than 1%) of these cells expressed all the characteristic "mature" markers. Limit-dilution precursor frequency studies showed the level of functional cells in the outer cortex was extremely low. The overall conclusion was that the vast majority of cells of complete "mature" phenotype are confined to the thymic medulla. These findings favor the view that thymus migrants originate from the thymic medulla, but do not exclude a cortical origin. The results also illustrate the need for multiparameter analysis to distinguish medullary thymocytes from early blast cells.     

The distribution of adenosine deaminase among lymphocyte populations in the rat.

Adenosine deaminase (ADA) activity was determined in young rat lymphocyte populations. The ADA-specific activity (per 10(8) cells and per milligram protein) was 3- to 10-fold higher in thymocytes than in lymphocytes from thoracic duct, lymph node, spleen, and bone marrow. The high ADA activity in thymocytes appeared to be preferentially associated with cortical thymocytes. Enrichment or depletion of cortical thymocytes by density gradient centrifugation, cortisone treatment, or selective lysis with anti-Thy-1 plus complement resulted in parallel increases or decreases in ADA levles. These results also suggested that medullary thymocytes have ADA levels similar to those of peripheral lymphocytes. "Immature" cortical thymocytes and thymocyte progenitors appeared to have low ADA activity; low enzyme levels were found in fetal thymus at 16 days of embryonic life, in the early phases of thymus regeneration, and in a "null" cell population isolated from bone marrow. This study demonstrates that ADA activity varies markedly during T lymphocyte differentiation and suggests that fundamental differences in nucleotide metabolism may exist in T cells at different stages of development.     

Ontogeny of murine T lymphocytes: I. Maturation of thymocytes induced in vitro by tumor necrosis factor-positive serum (TNF+)

One question which is unresolved in developmental immunology is whether cortical thymocytes are the precursor cells which give rise to medullary thymocytes and peripheral T cells. Cortical thymocytes display a characteristic surface antigen phenotype (high TL and Thy-1, low H-2, no Qa-2, no Qa-3), are agglutinated by peanut agglutinin (PNA), and are unresponsive to concanavalin A (Con-A). The functionally more mature medullary thymocytes express a surface phenotype more closely resembling peripheral T cells (no TL, low Thy-1, high H-2, and some Qa-2), are not agglutinated by PNA, and are responsive to Con-A. An in vitro induction system has been devised in which mouse thymocytes undergo quantitative changes in surface antigens in less than 24 hr and increase their mitogen response to Con-A. The phenotypic changes are characterized by a decrease of TL and Thy-1 and an increase in H-2, Qa-2, and Qa-3. Studies in which thymocytes were fractionated on BSA gradients and by PNA agglutination demonstrate that the inducible cells have the properties of cortical thymocytes. Our data show that a subpopulation of cortical thymocytes can acquire phenotypic characteristics similar to medullary thymocytes and peripheral T cells.     

1,25-Dihydroxyvitamin D3 inhibits selectively the mitogenic stimulation of mouse medullary thymocytes

Mouse thymocytes were separated into cortical and medullary subpopulations by differential agglutination with peanut agglutinin. A high-affinity receptor for 1,25-dihydroxyvitamin D3 is present in medullary immunocompetent mouse thymocytes and is absent from cortical immature cells. 1,25-dihydroxyvitamin D3, at physiological concentrations, inhibits the mitogenic response of the medullary cells to phytohemagglutinin and interleukin-2, but has no effect on the cortical subpopulation. Other less active metabolites of vitamin D had little or no effect on medullary cell stimulation.     

T cell maturation: thymocyte and thymus migrant subpopulations defined with monoclonal antibodies to MHC region antigens

The maturation sequences of thymocytes is known to some extent: A generative layer of subcapsular large lymphoblasts gives rise to a major population of small cortical thymocytes and a minor population of midsize medullary thymocytes. The relative contribution of these three populations to the peripheral T cell populations is not yet known. In this study, subcapsular lymphoblasts, cortical small cells, medullary cells, and thymic emigrant cells have all been analyzed by immunofluorescence for expression of the antigens H-2D, I-A, H-2K, and TL. H-2D is expressed brightly on all subcapsular large cells, dimly on cortical small cells, and brightly on all migrants, cortisone-resistant thymocytes (CRT), and peripheral T cells. I-A can be detected at low levels on 30 to 50% of cells in all the thymic subpopulations, and on 30 to 50% of migrants and peripheral T cells. Fifty to 80% of small cortical cells do not express detectable H-2K, but all the other subpopulations, both inside and outside the thymus, stain uniformly quite brightly. TL3 is expressed on 70 to 80% of subcapsular and cortical thymocytes, 30 to 40% of CRT, is undetectable on migrants but can be seen at low levels on 10 to 20% of spleen and lymph node T cells. The possibility that some or all of these antigens represent stable markers of separate lineages rather than unstable, stage-specific markers is discussed.     

The distribution of terminal deoxynucleotidyl transferase (TdT) among subsets of thymocytes in the rat.

Terminal deoxynucleotidyl transferase (TdT), a unique DNA-polymerizing enzyme,has been shown to be present in a moderately dense subpopulation of rat thymocytes separated on discontinuous Ficoll density gradients. This subpopulation has been characterized by using antigenic and functional markers to identify directly and quantify cortical and medullary thymocytes. The TdT-positive thymocytes are depleted by cortisone administration, lack responsiveness to phytohemagglutinin, concanavalin-A, and histocompatibility alloantigens, bear surface antigens characteristic of cortical thymocytes (bone marrow lymphocyte antigen) and lack surface antigens characteristic of medullary thymocytes (rat-masked thymocyte antigen and histocompatibility antigens). The results indicate that TdT is present exclusively (or in markedly higher concentrations) in a subset of cells which comprised about 65% of cortical thymocytes. Two other major subsets of cortical thymocytes were identified which appeared to be TdT-negative. A minor subset of very low density cortical thymocytes was also defined. These observations have provided insight into the possible pathways of thymocyte ontogeny.     

Fine structure of small lymphocytes in the thymus of the mouse: Qualitative and quantitative analysis by electron microscopy

Summary Thymic small lymphocytes of dd-mice were qualitatively and quantitatively studied by electron microscopy. Differences in fine structure were revealed between cortical and medullary small lymphocytes. Cortical small lymphocytes are rounded in cell outline with a round nucleus. The cytoplasm surrounding the nucleus as a narrow rim is scanty and appears relatively dense due to an abundance of free ribosomes. The cell organelles are not well developed. On the other hand, medullary small lymphocytes are more irregular in shape with uneven cell membranes. Their nuclei are also more irregular in outline with frequent infoldings of the nuclear membrane. The cytoplasm is more abundant and paler with less numerous ribosomes. The cell organelles are better developed.Quantitative analysis was made of both cortical and medullary small lymphocytes by means of the point counting method. The nuclei of both cortical and medullary small lymphocytes are almost the same in size (a diameter of 4.9 ). The cell sizes are different between cortical and medullary lymphocytes: cortical small lymphocytes with a diameter of 5.5 were smaller than medullary ones with a diameter of 6.4 .Cortical small lymphocytes are very sensitive to the destructive effects of hydrocortisone, whereas the medullary ones are resistant.Periarteriolar lymphoid sheath in the splenic white pulp, which is known to be a thymus-dependent area, contains small lymphocytes that were similar in cytological details to medullary small lymphocytes of the thymus.In the light of the recent knowledge about a recirculating long-lived small lymphocyte pool, it appears probable that medullary small lymphocytes represent a contribution to the pool and that small lymphocytes with a long life span can be cytologically identified by electron microscopy.     

A homing receptor-bearing cortical thymocyte subset: implications for thymus cell migration and the nature of cortisone-resistant thymocytes

The thymus exports a selected subset of virgin T lymphocytes to the peripheral lymphoid organs. The mature phenotype of these thymus emigrants is similar to that of medullary thymocytes and has been cited as supporting a medullary rather than cortical exit site. Using the monoclonal antibody MEL-14, we identify a 1%-3% subpopulation of thymocytes that expresses high levels of a receptor molecule involved in lymphocyte homing to peripheral lymph nodes. We present evidence that these rare MEL-14hi thymocytes are predominantly of mature phenotype and represent the major source of thymus emigrants. Surprisingly, MEL-14hi thymocytes are exclusively cortical in location, although their mature phenotype may allow them to masquerade as medullary cells in conventional studies. We also demonstrate that unlike medullary thymocytes, many cortisone-resistant thymocytes (CRT) are MEL-14hi. Thus, in contrast to current dogma, CRT do not represent a sample of medullary thymocytes as they are found in situ and their level of immunocompetence does not necessarily reflect that of the medullary population. Our findings refute the hypothesis that phenotypically and functionally mature cells are restricted to the medulla, and support our proposition that most thymus emigrants are derived from the MEL-14hi cortical subset.     

Relationship between structure and T-lymphocyte maturation in human thymomas. Enzyme histochemical and immunohistological studies

Twenty-six human thymomas were studied in an attempt to correlate their morphological appearance with the type and degree of T-lymphocyte maturation, as determined by acid alpha-naphthyl-acetate esterase (ANAE) activity and immunological analysis. Four normal human thymuses were used for purposes of comparison. Two morphological patterns were identified in the thymomas. The distinction was based largely on similarities between the neoplastic epithelial cells and normal cortical and medullary epithelial cells, and on the relative proportions of epithelial cells and lymphocytes. By these criteria "medullary" and "cortical" patterns were identified. In several thymomas both patterns were present in the same tumor ("mixed-type pattern"), producing alternating dark cortical-like areas and lighter foci of medullary differentiation. A good correlation was found between the two patterns and the phenotype of the T-associated lymphoid component. ANAE activity, which was completely lacking in normal cortical thymocytes, was almost absent in the phenotypically immature T-cells of cortical-type thymomas. By contrast, in the medullary-type thymomas, T-cells showed immunological features in common with medullary thymocytes. This was characterized by strong ANAE activity in the majority of cells with a staining pattern corresponding to that of peripheral T-lymphocytes. In addition, most of the proliferating epithelial cells in medullary-type thymomas stained strongly with anti-cytokeratin and anti-epidermal-type keratin antisera. In the mixed-type thymomas the epithelial cell morphology and the immunohistochemical and enzymic features of the T-cells were found to be closely related to the respective cortical--or medullary-like areas. It was concluded that the various characteristics of normal thymic cortex and medulla studied are also present in thymomas. In particular, in medullar-type thymomas the presence of many of the features of normal thymic medulla, such as a squamous cell component, macrophages and interdigitating reticulum cells, may constitute a microenvironment which operates actively in T-cell education. This may account for the functional activities, characteristic of peripheral T-lymphocytes, which T-lymphocytes attain in these thymomas.     

Immunohistochemical characterization of nurse cells in normal human thymus.

Lymphoepithelial complexes known as thymic "nurse" cells (TNC) have been isolated and described in the thymus of several animal species including man. Most of the investigations on TNC have been carried out in enzymatically digested thymuses in which TNC were isolated by differential sedimentation. In the present study we demonstrate TNC in immunohistochemically stained sections of human thymus as ring-shaped cells completely enclosing thymocytes and localized not only in the cortex, but also at the corticomedullary junction where they have not been previously described. TNC expressed epithelial markers [low and high molecular weight keratins identified by 35 beta H11 and 34 beta E12 monoclonal antibodies, a cortical antigen shared with neuroectodermal neoplasms recognized by the GE2 monoclonal antibody, and tissue polypeptide antigen (TPA:B1)], class II histocompatibility antigens (HLA-DR), and thymosin alpha 1. Double staining experiments with the nuclear proliferation-associated antigen Ki-67 and the cortical epithelium marker GE2 showed that most thymocytes enclosed in these cortical TNC were not proliferating. The antigens expressed by TNC indicate that not only cortical, but also medullary epithelial cells are part of the TNC system. The possible role of TNC in the education and maturation of thymocytes is discussed.     

Thymic epithelium in vitro. V. Binding of thymocytes to cultured thymic epithelial cells

Direct contact between thymocytes and thymic stromal elements may be one of the mechanisms involved in thymocyte differentiation. Thymic lymphoepithelial complexes have been isolated in which thymocytes appear to be in direct association with cortical epithelial cells. We have previously reported the isolation and successful culture of two morphologically distinct types of murine thymic epithelial cells. We have utilized these to study the interactions of lymphoid and epithelial cells by means of an in vitro assay of the binding of radiolabeled thymocytes to monolayers of these cultured thymic epithelial cells. The percentage of bound cells increased rapidly during the first hour of incubation, reaching approximately 40% binding. Binding continued to increase slowly until plateau levels were reached at approximately 5 hr. Thymocyte binding to thymic epithelium, but not fibroblast monolayers, was trypsin-sensitive, suggesting that specific protein interactions may be involved. Binding of thymocytes to epithelium was temperature-dependent, involved formation of cytoplasmic projections, and was inhibited by cytochalasin B. We also found that cortical thymocytes (peanut agglutinin-positive (PNA+)cells) bound to cultured epithelium to a greater degree than medullary thymocytes (PNA- cells). This correlates with in vivo studies by others in which thymocytes associated with lymphoepithelial complexes have been found to have immature phenotypes. This system provides a means for a quantitative study of the role of cell to cell contact in the process of thymocyte selection and differentiation.     

In situ localization of CD45 isoforms in the human thymus indicates a medullary location for the thymic generative lineage.

Previous work has suggested that the generative lineage within the human thymus can be defined by the selective expression of CD45 isoforms and is CD45RO- and predominantly CD45RA+. In order to physically localize these cells we have stained frozen sections of human thymus with antibodies to CD45RO (p180), and CD45RA (p205/P220), as well as with CD1 and HLA class I to define cortical and medullary areas, respectively. In the cortex, 70 to 90% of thymocytes were CD45RO+, whereas only 0.5% expressed CD45RA. Medullary cells were 30% CD45RO+, 29% CD45RA+; approximately 40% did not express detectable levels of either isoform but did express CD45 common determinants. To assess the degree of proliferation of cells expressing CD45 isoforms, we stained adjacent sections, or used double staining, with Ki67, an antibody that detects a nuclear Ag on proliferating cells. We found that CD45RA+ thymocytes are predominantly a resting medullary population with a small component in cell cycle, consistent with our analysis of human thymocytes by immunofluorescence, and with data in murine systems defining the generative lineage. To confirm that the CD1- or low, CD45RO-CD45RA+ thymocytes defined by immunofluorescence analysis were likely to have a medullary location, we analyzed the CD4/CD8 subset distribution of CD1-cells. From 80 to 90% of CD1-thymocytes are CD4+ or CD8+ single positives or CD-8- double negatives. CD1-thymocytes also include 12 to 14% CD4+8+ cells with a probable medullary location. A similar analysis of lymphocytes expressing a high density of HLA class I, which have a medullary location, confirmed the existence of CD4+8+ thymocytes in the medulla. Purified CD3-4-8- cells, previously shown to be CD1-CD45RA+, were also shown to bear a high density of HLA class I, indicating a medullary location. Correlative localization of a panel of Ag thus supports the argument for a medullary location of the thymic generative lineage.     

High-level secretion of interleukin 2 by a subset of proliferating thymic lymphoblasts

We have characterized the thymocytes that can be induced to secrete interleukin 2 (IL 2) after polyclonal stimulation with Con A. For maximal activation, an important adjunct to the Con A is the phorbol ester TPA. In the presence of TPA, IL 2 production by thymocytes is relatively independent of adherent accessory cells; this allows us to compare the abilities of different thymic subpopulations to make IL 2. The most numerous class that includes IL 2 producers is made up of cells with a typical "medullary" population, the phenotype: moderately small, postmitotic cells that fail to bind peanut agglutinin. In addition, however, a population of large, proliferating lymphoblasts is competent in IL 2 production directly as isolated. Relative to the total "medullary" population, the lymphoblasts are enriched for the ability to make IL 2. They account for a significant proportion of the total IL 2 produced by thymocytes, and demonstrate that this aspect of immunocompetence is not restricted to cells that have finished their intrathymic proliferation. The IL 2-producing lymphoblasts do not bind peanut agglutinin or express thymus-leukemia antigen, but they do express high levels of Lyt-1. Although distinct from most medullary thymocytes, therefore, they are also distinct from the majority of cortical blast cells for which a direct precursor role has been established. They may be a subset of the rare proliferating blast cells in the medulla. Further heterogeneity in the thymic IL 2 producers is demonstrated by their expression of the Lyt-2 glycoprotein. The majority of IL 2 producers are Lyt-2- as are the majority of peripheral T "helper" cells. However, a distinct minority of the thymic IL 2 producers express Lyt-2. Therefore, the ability of some peripheral Lyt-2+ cells to secrete IL 2 may be determined at the time of their initial programming in the thymus.     

Intrathymic differentiation of cytotoxic T lymphocyte (CTL) precursors. I. The CTL immunocompetence of peanut agglutinin-positive (cortical) and negative (medullary) Lyt 123 thymocytes

To analyze the developmental and functional interrelationship between cortical and medullary thymocytes, the peanut agglutinin-(PNA) binding capacity was used to separate thymocytes into PNA+ (cortical) and PNA- (medullary) thymocytes. Virtually, all positively selected PNA+ thymocytes (90% of the overall thymocyte population) expressed the Lyt 123 phenotype, whereas 90% of negatively selected PNA- thymocytes expressed Lyt 1 alloantigens, about 10% being Lyt 123 thymocytes. Provided, the requirement of Lyt 1 T helper cells was bypassed by Interleukin 2, a nonspecific mediator of T help, PNA+ Lyt 123 thymocytes mounted cytotoxic T cell responses comparable in magnitude to that of peripheral T cells. Their repertoire included antigenic disparities coded for by the complete MHC complex, H-2K, I-A, H-2D, mutational events at H-2K, as well as antigenic disparities expressed on TNP conjugated- and Sendai virus-infected syngeneic cells. PNA- Lyt 123 thymocytes represent a highly reactive pool of primary cytotoxic T lymphocyte (CTL) precursors for both alloreactive and H-2-restricted CTL responses. Since PNA- thymocytes include also Lyt 1 T helper cells, PNA- responder thymocytes are able to mount autonomously (CTL responses. Our data are first to provide direct evidence that Lyt 123 cells represent a common source of alloreactive and H-2-restricted CTL precursors in unprimed lymphocyte populations. Moreover, the apparent immunocompetence of cortical PNA+ thymocytes is now explained by their lack of T helper cells.     

Differentiation of human T lymphocytes. I. Acquisition of a novel human cell surface protein (p80) during normal intrathymic T cell maturation

The thymus is thought to be the primary central lymphoid organ in which T cells mature. Although thymic cortical and medullary compartments are distinct histologically, few antigens have been described that are absolutely acquired during the presumed intrathymic maturation pathway from cortical to medullary thymocytes. In this paper, we describe the acquisition during human intrathymic T cell maturation of a novel protein (p80) defined by a monoclonal antibody (A1G3). Although the p80-A1G3 antigen is distributed throughout the body and is not T cell specific, our study demonstrates that expression of p80-A1G3 antigen in normal human thymus is associated with thymocyte functional maturity and location in the thymus medulla. Moreover, in contrast to other markers of mature human T cells, the p80-A1G3 cell surface protein is not expressed on T6+ cortical thymocytes, and, therefore, is absolutely acquired by medullary thymocytes during T cell maturation. Thus, the p80-A1G3 antigen and the A1G3 antibody provide a heretofore unavailable system for the study of molecular events that transpire during the maturation of thymocytes.     

Precursor immigration and thymocyte succession during larval development and metamorphosis in Xenopus

The developing thymus in Xenopus was examined at four different levels: 1) precursor immigration of cytogenetically distinct embryonic stem cells; 2) waves of colonization during tadpole life and metamorphosis; 3) inter-thymic exchange of cells between separate lobes; and 4) development of cortical and medullary thymocytes. Based on the flow cytometric analysis of cytogenetically distinct thymocytes, there were at least two periods of stem cell immigration into the thymus, one during early larval life and the second before or during metamorphosis. Within the thymus, cohorts of cells derived from the first wave of immigration expanded at different times. The initial expansion occurred before 35 days of development. Cells involved in the second period of expansion were also derived from the initial immigrants, expanded after 35 days, and resulted in a turnover of thymocytes during the larval period. Precursor cells entering the thymus during metamorphosis expanded and resulted in an additional replacement of thymocytes. Cortical and medullary thymocytes were isolated from animals that received embryonic stem cell grafts. No differences in the presence or absence, or in the percentages, of donor thymocytes in these different fractions were observed. When limiting numbers of stem cells were transplanted, several cases of asymmetrical thymic lobe colonization were observed. These data suggested that an inter-thymic exchange of cells did not occur during larval life.     

Differential expression and regulation of cyclooxygenase isozymes in thymic stromal cells

Prostaglandins (PGs) are lipid-derived mediators of rapid and localized cellular responses. Given the role of PG in supporting thymic T cell development, we investigated the expression of the PG synthases, also known as cyclooxygenases (COX)-1 and -2, in the biosynthesis of PGs in thymic stromal cell lines. The predominant isozyme expressed in cortical thymic epithelial cells was COX-1, while COX-2 predominated in the medulla. IFN-gamma up-regulated expression and activity of COX-2 in medullary cells, in which COX-2 was expressed constitutively. In contrast, IFN-gamma down-regulated COX-1 activity, but not expression, in cortical cells. Stromal cells support T cell development in the thymus, although the mediators of this effect are unknown. Selective inhibition of COX-2, but not COX-1, blocked the adhesion of CD4+CD8+ and CD4+CD8- thymocytes to medullary cell lines. No effect of the inhibitors was observed on the interactions of thymocytes with cortical epithelial lines. These data further support the differential regulation of COX-1 and COX-2 expression and function in thymic stromal cells. PGs produced by COX-2 in the medullary thymic stroma may regulate the development of thymocytes by modulating their interaction with stromal cells.     

Dual immunofluorescence studies of cortisone-induced thymic involution: evidence for a major cortical component to cortisone-resistant thymocytes

Cortisone-resistant thymocytes (CRT) have been used as the experimental equivalent of medullary thymocytes for the past 15 yr. Studies with CRT have provided evidence that the medullary population is similar to mature T cells in phenotype and function and may therefore be the major source of thymus emigrants. However, we have recently demonstrated that CRT differ from medullary thymocytes in their expression of the homing receptor molecule recognized by the monoclonal antibody MEL-14. Thus, many CRT express high levels of the MEL-14-defined homing receptor, whereas medullary thymocytes are MEL-14- to MEL-14lo. In normal adult mice, only 1 to 3% of thymocytes are MEL-14hi; these cells are located exclusively in the cortex and many are phenotypically and functionally mature. In this study we have used dual immunofluorescence techniques to further characterize those thymocytes resistant to cortisone treatment. Aside from being of mature phenotype with respect to expression of peanut agglutinin binding sites and the cell surface molecules H-2K, Ly-1, Lyt-2, and L3T4, CRT can be divided into MEL-14lo and MEL-14hi subpopulations, suggesting that they may actually be derived from both the medullary and the MEL-14hi cortical thymocyte subsets.     

Distribution of thymocytes expressing gamma delta receptors in the murine thymus during development

We have examined the appearance of thymocytes expressing gamma delta TCR within the developing thymus by using immunohistochemical techniques and flow cytometry in conjunction with the mAb 3A10, which recognizes a determinant associated with the constant region of the delta-chain. gamma delta+ Cells were first detected at day 16 of gestation, attained maximal levels at day 17 of gestation, and declined thereafter. By using the Ulex europeus agglutinin to identify medullary epithelial cells in situ, we observed a striking colocalization of gamma delta+ thymocytes and U. europeus agglutinin-positive medullary epithelial cells during late fetal and neonatal periods of development. In the thymuses of adult mice, gamma delta+ thymocytes were scattered throughout cortical and medullary areas of the thymus and most concentrated in the subcapsular areas of the thymus. Ultrastructural immunohistochemistry confirmed the close association between medullary thymic epithelial cells and gamma delta+ thymocytes in the neonatal thymus and also showed that some TCR-gamma delta molecules were patched to areas of contact with medullary epithelial cells. In contrast to the cellular distribution of either CD3 molecules or the TCR-alpha beta, where extensive intracellular labeling of thymocytes has been observed, cytoplasmic accumulation of delta-chain was not detected.